The development of multi-drug resistant Staphylococcus infections is an increasing concern for the global health community. Infection caused by methicillin-resistant Staphylococcus aureus (MRSA) is becoming increasingly difficult to treat with conventional antibiotics, leading to a sharp rise in clinical complications.
The gram-positive microbe, Bacillus anthracis, is the etiologic agent of anthrax, a disease common to livestock but also considered to be a primary biological warfare and bioterrorism threat to humans worldwide. Humans can be infected by exposure to spores or to infected animals or their waste products. Prompt medical attention of anthrax infections is required, and typically effective, by treatment with penicillin G, doxycycline, or the quinolone ciprofloxacin. If untreated, death usually results within several days to a week. There is growing concern that B. anthracis is developing resistance to the common drugs that are used to treat the disease, and new antibiotics possessing a different mode of action are urgently needed.
The discovery of N-alkylthio β-lactams in the Turos group has opened a new pathway in the development of potent anti-infectives (Turos, E. et al. Bioorg. Med. Chem., 2005, 13:6289-6308; Long, T. E. et al. Bioorg. Med. Chem., 2003, 11:1859-1863; Turos, E. et al. Bioorg. Med. Chem. Lett., 2002, 12:2229-2231). While the initial results have been exciting, much of their mode of action remains to be completely understood.
Long (“N-thiolated β-lactams: Chemistry and biology of a novel class of antimicrobial agents for MRSA” University of South Florida Dissertation, 2003) and Heldreth (“N-thiolated beta-lactams: Chemistry, SAR and intracellular target of a novel class of antimicrobial and anticancer agents” University of South Florida Dissertation, 2004) were able to show that a disulfide-forming thiol transfer to intracellular thiols such as coenzyme A or glutathione was a key metabolic pathway for this class of compounds. Further, it was demonstrated that these drugs are most effective in those strains of bacteria where coenzyme A is present in high concentrations relative to other cytosolic thiols (Long, T. “N-thiolated β-lactams: Chemistry and biology of a novel class of antimicrobial agents for MRSA” University of South Florida Dissertation, 2003; Newton, G. L. et al. J. Bacteriology, 1996, 178(7):1990-1995), and, thus, where mixed CoA disulfides are produced within the cells (see scheme of thiol transfer from β-lactam to CoA in FIG. 1A).
Although the abundance of CoA in Staphylococcus and Bacillus stems from its role within the thiol-redox buffer of these strains (Kaleidagraph, version 3.5.1.0, available from Synergy Software), the thiol-redox buffer is not directly affected by reaction of the N-alkylthio β-lactams with CoA. Although CoA disulfide reductase was ruled out as an enzymatic target, at least one of the two key mechanistic postulates is likely still valid:
1.) The CoA mixed disulfide produced by reaction of CoA with the N-alkylthio β-lactams acts as an inhibitor of a CoA-selective enzyme.
2.) The N-alkylthio β-lactams are actually the active molecules, and the thiol-transfer reaction of CoA functions as a natural defense mechanism.
If postulate (1) is correct, it stands to reason that a large number of enzymatic pathways may be affected, since CoA is reported to be involved with approximately four percent of all known enzymatic reactions (Lee, C-H. and Chen, A. F. “Immobilized coenzymes and derivatives” in The Pyridine Nucleotide Coenzymes, Everse, J. et al. Eds, New York: Academic Press, 1982). Nonetheless, the elucidation of individual pathways is important to the further development of these valuable compounds.
N-Alkylthioβ-Lactams and Fatty Acid Synthesis
Recent experiments conducted by Dr. Seyoung Jang of the Turos group demonstrated that treatment of S. aureus with lactam 8 (FIG. 2) greatly reduced uptake of radiolabeled acetate (as shown by the graph in FIG. 3) (Leonardi, R. et al. J. Biol. Chem., 2005, 280:3314-3322; Slater-Radosti, C. et al. Antimicrobial Chemother., 2001, 48:1-6; Higgins, D. L. et al. Antimicro. Agents Chemother., 2005, 49(3):1127-1134; Bligh, E. G. and Dyer, W. J. Can. J. Biochem. Phys., 1959, 37:911-197). This reduction in uptake suggests that a key mechanistic pathway of the N-alkylthio β-lactams is the inhibition of fatty acid synthesis.
Bacterial fatty acid synthesis, often called Type II FAS, has been a target for antibiotics development for nearly a decade (Marrakchi, H. et al. Biochem. Sco. Trans., 2002, 30(6):1050-1055; Campbell, J. W. and Cronan, J. E. Ann. Rev. Microbiol., 2001, 55:305-352; Cronan, J. E., Jr. and Rock, C. O. in E. Coli and Salmonella typhimurium: Cellular and Molecular Biology (Neidhardt, F. C. et al. Eds.), Washington, D.C.: American Society for Microbiology, 1996; Rock, C. O. and Cronan, J. E., Jr. Biochim. Biophys. Acta., 1996, 1302:1-16; Heath, R. J. and Rock, C. O. J. Biol. Chem., 1996, 271(18):10996-11000; Daines, R. A. et al. J. Med. Chem., 2003, 46:5-8; Jones, P. B. et al. J. Med. Chem., 2000, 43:3304-3314; He, X. and Reynolds, K. A. Antimicrob. Agents Chemother., 2002, 46(5):1310-1318; Jones, P. B. et al. J. Med. Chem., 2000, 43:3304-3314; Nie, Z. et al. J. Med. Chem., 2005, 48:1596-1609; He, X. et al. Antimicro. Agents Chemother., 2004, 48:3093-3102; Choi, K-H. et al. J. Bacteriology, 2000, 182(2):365-370). Unlike mammalian (Type I) FAS, which operates through a multienzyme complex (Voet, D. and Voet, G. Biochemistry, 3rd Ed., John Wiley and Sons: Hoboken, N.J., 2004), bacterial FAS utilizes a series of discreet enzymes. This key difference opens opportunities for the development of inhibitors which are selective for Type II FAS (Marrakchi, H. et al. Biochem. Sco. Trans., 2002, 30(6):1050-1055).
Type II FAS is depicted in the scheme of FIG. 4 (Marrakchi, H. et al. Biochem. Sco. Trans., 2002, 30(6):1050-1055; Voet, D. and Voet, G. Biochemistry, 3rd Ed., John Wiley and Sons: Hoboken, N.J., 2004). The synthetic cycle represents a recurring sequence of condensation (two-carbon elongation), reduction, dehydration, and reduction. Throughout the fatty acid cycle, the acyl substrates are carried from enzyme to enzyme by the acyl carrier protein (FIG. 5). This small protein contains a pantethine/thiol arm identical to that of CoA. Within this cycle, only two enzymes, malonyl/acetyl-CoA-ACP transacylase (MAT) and β-ketoacyl-ACP synthase III (FabH) utilize a CoA derivative.
Of the two possible enzymes, FabH seemed the logical first choice for exploration: first because of its role as the initial condensing enzyme in the FAS cascade, and secondly because FabH has already been purified from several species and its mode of action extensively studied (Marrakchi, H. et al. Biochem. Sco. Trans., 2002, 30(6):1050-1055; He, X. and Reynolds, K. A. Antimicrob. Agents Chemother., 2002, 46(5):1310-1318; Qui, X. et al. Protein Sci., 2005, 14:2087-2094; Davies, C. et al. Structure, 2000, 8:185-195; Scarsdale, J. N. et al. J. Biol. Chem., 2001, 276(23):20516-20522; Li, Y. et al. J. Bacteriol., 2004, 187:3795-3799).
The FabH active site contains a single cysteine residue located within a lipophilic pocket. Following binding of acetyl CoA, acetyl transfer from CoA to the active-site cysteine produces an S-acetyl cysteine and free CoA. Malonyl ACP then binds to the enzyme and undergoes a condensation reaction with the S-acetyl cysteine, with concurrent loss of CO2. Elimination of the cysteine produces β-ketobutanoyl-ACP, which is carried into the rest of the fatty acid cycle (FIG. 6).
Based on this mechanism, the inhibition of FabH by a CoA mixed disulfide was visualized. It was hypothesized by the present inventors that this inhibition involves not only a noncovalent inhibition by occupation of the binding pocket, but also a potentially irreversible inhibition via transfer of the alkylthio moiety to the active site cysteine of the FabH enzyme (FIGS. 1A-1B).
In order to explore these possibilities, it was proposed that a set of CoA mixed disulfides be tested for inhibitory activity against FabH. And, since it was possible that the N-alkylthio β-lactams might interact directly with this enzyme, compounds 2a and 2g (FIGS. 7A and 7B, respectively) were also tested against FabH.
For the past decade, there has been considerable interest in the development of FabH inhibitors (Marrakchi, H. et al. Biochem. Sco. Trans., 2002, 30(6):1050-1055; Campbell, J. W. and Cronan, J. E. Ann. Rev. Microbiol., 2001, 55:305-352; Daines, R. A. et al. J. Med. Chem., 2003, 46:5-8; Qui, X. et al., Protein Science, 2005, 14:2087-2094; Jones, P. B. et al. J. Med. Chem., 2000, 43:3304-3314; He, X. and Reynolds, K. A. Antimicrob. Agents Chemother., 2002, 46(5):1310-1318). The key role of this enzyme in Type II fatty acid synthesis, as well as the differences between bacterial and mammalian FAS, make FabH a desirable target for the inhibition of bacterial growth (Marrakchi, H. et al. Biochem. Sco. Trans., 2002, 30(6):1050-1055; Campbell, J. W. and Cronan, J. E. Ann. Rev. Microbiol., 2001, 55:305-352).
It has recently reported that compounds such as the N-alkylthio β-lactams (FIG. 78A) (Turos, E. et al. Bioorg. Med. Chem. Lett., 2002, 12:2229-2231; Turos, E. et al. Tetrahedron, 2000, 56:5571-5578; Turos, E. et al. Bioorg. Med. Chem., 2005, 13:6289-6308; Coates, C. et al. Bioorg. Med. Chem., 2003, 11:193-196; Long, T. E. et al. Bioorg. Med. Chem., 2003, 11:1859-1863; Turos, E. et al. Bioorg. Med. Chem. Lett., 2002, 12:2229-2231; Heldreth, B. et al. Bioorg. and Med. Chem., 2006, 14:3775-3784) and N-alkylthio-2-oxazolidinones (FIG. 78B) (Mishra, R. K. et al. Bioorg. Med. Chem. Lett., 2006, 16(8):2081-2083) are able to inhibit the proliferation of methicillin-resistant Staphylococcus aureus (MRSA) and Bacillus anthracis. A major pathway for this inhibition involves thiol-transfer to Coenzyme A to produce an alkyl-CoA mixed disulfide (5 in the scheme of FIG. 79). Experimental evidence suggests that the resulting aryl-alkyl disulfide may inhibit FabH (6 in the scheme of FIG. 79) by reversibly “capping” the active site cysteine through a thiol-disulfide exchange (FIG. 79) (Revell, K. D. et al. Bioorg. and Med. Chem., 2007, 15(6):2453-2467).
In light of this, the semi-irreversible inhibition of FabH by CoA/alkyl mixed disulfides opens a new strategy for the blocking of this important enzyme.
Although alkyl-CoA disulfides have been shown to bind tightly to purified FabH, treatment of S. aureus cultures with these disulfides does not inhibit bacterial growth, likely because the multicharged CoA cannot easily traverse the cell membrane (Clarke, K. M. et al. J. Am. Chem. Soc., 2005, 237:11234-11235). However, it was postulated that smaller, non-ionic disulfides might be able to function as CoA disulfide mimics: they could block the FabH active site via thiol transfer in the same manner as the CoA mixed disulfides, but their smaller size and greater lipophilicity make it easier for them to enter the cell and inhibit FabH. To evaluate this hypothesis, a set of simple disulfides was prepared, and their activities against several bacteria were tested.
The design of CoA mixed-disulfide mimics was aided by previous research into FabH inhibitors. Jones et al. (Johns Hopkins University) reported that compound 30 (FIG. 8B) showed good activity against M. tuberculosis, presumably through inhibition of FabH (He, X. and Reynolds, K. A. Antimicrob. Agents Chemother., 2002, 46(5):1310-1318). Kevin Reynolds (Portland State) reported compound 31 (FIG. 8C) inhibits E. coli FabH. Researchers at GlaxoSmithKline recently published data on a class of FabH inhibitors, including the crystal structure of S. aureus FabH (Daines, R. A. et al. J. Med. Chem., 2003, 46:5-8), and the co-crystal structure of compound 29 (FIG. 8A, where R═(CH2)4CO2H) in E. coli FabH (Qui, X. et al. Protein Science, 2005, 14:2087-2094).
The crystal structures reported for compound 29 (Qui, X. et al. Protein Sci., 2005, 14:2087-2094) indicate that the active site is located within a fairly deep, lipophilic “pocket” into which acyl-CoA or acyl-ACP are able to stretch their lipophilic arm. Further, the co-crystal data reveals that, in the binding mode of compound 29, the 2,6-dichlorophenyl moiety reaches into the active site, such that the 4-position of the phenyl ring is proximal to the active site cysteine. Based on this, it was postulated that a molecule such as 32 (FIG. 9) might be able to effectively undergo thiol-disulfide exchange with the active site cysteine of FabH.
The present inventors postulated that, unlike noncovalent inhibitors, CoA disulfide mimics would not require long residence times in the active site. Rather, they need only to be able to enter the site and transfer the thiol. With this goal in mind, and in light of the efficacy of the nonpolar compounds 30 and 31, it was postulated that small, somewhat lipophilic disulfides of the general type (FIG. 80) would be best suited to access the pocket.
However, a deeper analysis suggests that compound 32 may be more specific than necessary. The effectiveness of benchmark compound 29 stems from its residence time in FabH. That is, its rate of binding to FabH is very fast, while its rate of dissociation is relatively slow.
In the case of a CoA mixed disulfide mimic, however, such long residence times may not be needed. Rather, it is critical that the disulfides be small enough to fit easily into the active-site pocket to transfer the thiol to the cysteine. To test this hypothesis, a set of very simple aryl-alkyl disulfides (FIGS. 10 and 11) were prepared by semicombinatorial methods and analyzed for activity against S. aureus, B. anthracis, and Escherichia coli. 